Vision and Weight Sensor Fusion for an Automated Hospital Waste Sorting System: A Practical Approach to Safer Healthcare Waste Management

Hospital waste management presents a critical chal- lenge in healthcare infrastructure worldwide, with improper segregation posing serious health and environmental risks. This research addresses the limitations of manual sorting and ex- isting automated systems by developing an intelligent waste segregation bin that combines computer vision with weight sensing. Our system utilizes a convolutional neural network (CNN) for visual classification of common hospital waste into four categories—plastic, glass, metal, and paper—while simulta- neously employing load cells for weight measurement. A novel fusion algorithm intelligently combines these complementary data streams, achieving a remarkable classification accuracy of 94.6%, significantly outperforming vision-only (88.2%) and weight-only (72.4%) approaches. The practical implementation includes a mechanical sorting mechanism and an IoT dashboard for real- time monitoring. Priced at approximately Rs.45,000 , our solution offers hospitals an affordable, reliable, and scalable alternative to error-prone manual sorting, potentially transforming waste management practices in healthcare facilities of varying sizes and resources.